The value of OSH: Estimating the societal costs of work-related injuries and diseases

Introduction

Insight into the financial consequences of work-related injuries and diseases provides governments, policy makers and employers’ organisations with relevant data for the purpose of developing OSH policies and agreements. Also, insight into these costs will help to raise awareness of the magnitude of the problem and will contribute to a more efficient allocation of resources for OSH. Different methods are developed to estimate these costs. A bottom-up model builds up from components of costs to total costs. In a top-down model the total costs are estimated via the total burden of injury and disease, and the estimated fraction that was caused by occupational factors. This article describes two models to estimate the costs of work/related injuries and disease, one following a bottom up model, one following a top/down model. Results of the models are shown of five European countries; Finland, Germany The Netherlands, Italy and Poland.

Bottom-up model

The bottom-up model described in this draws on methods synthesized from a number of studies, specifically the following: 1) a study on the economic burden of lung cancer and mesothelioma due to occupational and para-occupational asbestos exposure^[1]; 2) a study on the economic burden of occupation injury and disease in the United States^[2]; and 3) a study on the cost of work-related injury and disease for Australian employers, workers and the community^[3]. These studies are all based on the bottom-up model and use incident cases of work-related injuries and diseases as the starting point, and estimate the direct, indirect and intangible costs associated with these cases.

For the estimation of non-fatal work-related disease cases, different data sources were consulted leading to different scenarios of the case count. In the baseline scenario, we started with the count of compensated (accepted, recognized) and non-compensated (suspected) non-fatal cases for each country for most types of diseases, with the following exceptions: for cancers, circulatory diseases, respiratory diseases, and musculoskeletal diseases we estimated case counts from the database of the Burden of Disease (BoD) study as registered by the Institute for Health Metrics and Evaluation  (IHME), and used the attributable fraction derived from this database. We also defined a low limit scenario (i.e. only includes compensated cases), and a high limit scenario (i.e. all types of work-related disease estimated using attributable fractions). Data from the LFS ad hoc module 2013^[5] were used to estimate the distribution of the non-fatal work-related disease cases by age, as well as severity (number of workdays lost). Finally, the estimation of the fatal work-related diseases was also based on the IHME database and attributable fractions derived from this database. The figures presented in this summary are based on the baseline scenario.

In the model, three high-level cost categories are considered: direct costs, indirect costs and intangible costs. Direct costs include all healthcare related products and services, whether paid for by the public sector, insurer, employer, worker or other stakeholder. We focus on four direct cost items: 1) healthcare costs paid for by the public sector/insurer; 2) public sector/insurer administration/overhead costs; 3) informal caregiving time from family and community; and 4) worker out of pocket costs for healthcare products and services, including costs associated with seeking care. We estimated six key subcomponents of indirect costs: 1) market output losses due to absenteeism and reduced work ability associated with permanent impairment; 2) payroll/fringe benefits associated with wages and salaries; 3) employer adjustment costs, 4) insurance administration costs associated with disability insurance/workers’ compensation; 5) home production losses; and 6) presenteeism associated with paid employment activity. Finally, intangible costs refer to losses associated with health-related quality of life. Health-related quality of life is estimated in terms of Quality Adjusted Life Years (QALYs) and then monetized.

The cost estimations begin with incidence counts (cases) of work-related injuries and diseases to estimate the total costs in a particular cost category, which are then multiplied by the resources associated with being a case and a price weight if the measure of resources is in non-monetary units (e.g., months lost from paid employment due to work disability). Incidence counts have been stratified by sex, age bracket, type of injury (high-level ESAW categories), and severity (based on days absent from work). A representation of the formula is as follows:

Total (sub) category costs for a strata = # of cases in the strata x per case cost for the strata

The results are presented below. Table 1 shows the estimation of the number of cases in each country and Table 2 presents the estimates of the costs. Finally Table 3 presents the economic burden of work-related injury and disease by stakeholder.

Top down model

The top-down model in the present study is based on DALYs, i.e. Disability Adjusted Life Years. The DALY is a measure of overall disease burden, expressed as the number of healthy years lost due to early death or due to living with ill-health. DALYs are calculated by disease category and are the sum of life years lost due to premature mortality and ‘healthy’ life years lost due to disability. The latter is calculated by multiplying the number of cases by duration and the disease-specific disability weight. A disability weight is a weighting factor that reflects the severity of the disease on a scale from 0 (perfect health) to 1 (equivalent to death). The baseline variant in the present study is based on the DALYs by cause, sex, age and country taken from the World Health Organisation (WHO) Global Health Estimates: Global burden of disease estimates 2000-2016, as published by the WHO department of Information, Evidence and Research in June 2018^[6].

In the literature, three broad methodological approaches to estimating the monetary value of a DALY can be identified: 1) human capital approach, 2) willingness-to-pay approach and 3) value of statistical life (year) approach. In the human capital approach, the monetary value of a DALY is based on the loss of economic productivity due to ill health, disability, or premature mortality. A drawback of the human capital monetization approach is that only part of an individual’s welfare is measured. Life beyond paid work is not valued. Theoretically, the two other monetization approaches considered in this report, the willingness-to-pay and value of a statistical life year approach do include valuations for broader aspects of life. The willingness-to-pay (WTP) approach is based on preferences of survey respondents to pay for health gains. The value of statistical life (VSL) represents a total monetary value of an average adult towards the life expectancy age, hence a value for the total remaining lifetime of an average person in case of no injury or disease, which in fact is often also obtained with willingness-to-pay surveys. The drawback of both WTP and VSL approach is that values are based on surveys and valuation methods that are highly sensitive to the questions asked. As a result of the sensitivity to methods, the variance in values found across studies is quite wide. Variance in values is also wide in the human capital approach. For example, according to the recommendations of the WHO Commission on Macroeconomics and Health, the monetary indicator varies between GDP per capita and 3*GDP per capita (Harvard School of Public Health & World Economic Forum, 2011).

In conclusion, within each monetization approach, the range of monetary values as found in the literature were large. Therefore, we worked with the minimum, mean, median and maximum of these values in our models. Table 4 contains the results of the top-down model by country, according to different monetization approaches.

Results of both models compared

In the bottom-up model, the total estimated economic burden of work related injuries and diseases – including fatal and non-fatal cases– ranges from 2.9% of GDP in Finland to 10.2% in Poland. In the top-down model, the economic burden is highly dependent on the monetization approach. In the human capital approach, the work-related economic burden varies from 0.6% to 4.5%, dependent on the monetization method, with less variance between countries. In the WTP approach, percentages are higher and vary from 0.5% to 8.3%. The VSLY approach yields the highest values, with estimates of the economic burden of occupation injury and disease at 1.4% of GDP at the minimum and 27.7% at the maximum. In this approach, variance between countries is also higher. The approach that comes closest to the results of the bottom-up model is the VSLY approach if we consider the average or median value of the different studies. Also the rank ordering of countries in terms of magnitude of economic burden relative to their GDP is similar to the bottom-up model with the highest value for Poland (average 10.2% and median 7.2% of GDP) and the lowest value for Finland (average 4.5% and median 4.1% of GDP). The similarity between the VSLY approach from the top-down model and the bottom-up model may be explained by the inclusion of health and life impacts in the VSLY approach. Health and life impacts, in the bottom-up model described as ‘intangible costs’ are a substantial part of the total costs in the bottom-up model, varying from 20% to almost 51%.

In comparing the outcomes of the two cost estimation models, it is important to realize that they do not estimate identical phenomena. Although they both are invoked to provide estimates of the economic burden of work-related injury and disease, the components of these models are very different. The bottom-up model provides more detailed information for policy makers such as direct, indirect and intangible costs, as well as costs by stakeholder. However, the top-down model has the advantages that far less time is needed to construct the model and country and regional comparisons are easier since international harmonized sources can be used.

Country comparison

In comparing the countries, we see in most scenarios that the economic burden of work-related injury and disease is relatively high in Poland and Italy, compared to Germany, Finland and The Netherlands. In Poland, at least part of it may be explained by the sector structure. The workforce in Poland consists of relatively many persons working in agriculture or industry. Although the percentage of persons working in industry in Italy is above average, the explanation for the relatively high burden is less clear than in Poland. Partly, the relatively burden is attributable to the relatively large number of DALYs lost to occupational lung cancer. However, the main difference with the other countries under study, is the relatively large number of DALYs lost to injuries, ‘unintentional injuries’ as well as ‘transport injuries’.

Implications for future projects

As noted, in this project on the economic burden of the work-related injury and disease, countries were selected based on the expectation that they had sufficient data of good quality to enable the estimation. Still, data were often lacking, quality was poor, alternative sources had to be explored to come to a reasonable estimation. In particular for the bottom-up model, which consists of several components, the search for appropriate data was quite a challenge, particularly for formal healthcare costs. Therefore, the first step to enable a cost estimation of this sort in all European countries, would be to build up and harmonize the data collected in these countries. We will list the main issues. First of all, the count of work-related injuries and diseases should be improved for all economic burden estimation models, whether they are inputs for a bottom-up model or used to estimate DALYs. In the present projects it was not possible to base the bottom-up model on incident cases of work-related diseases from country reporting. But data on incident cases for injuries and diseases has to come from somewhere for both the top-down and bottom-up models. Ideally from reliable, country specific sources so that meaningful cross country comparisons can be made. If they are approximated through generic, international sources, then cross country comparison is less meaningful for both models. Also, country specific data on the healthcare costs of injuries and diseases appeared to be very difficult to obtain. Finally, it would be helpful to come to a consensus on the way to value life and health impacts for both the top-down and bottom-up models.